tree-wide: s/getpid()/lxc_raw_getpid()/g

[mirror_lxc.git] / src / lxc / cgroups / cgfs.c

/*
 * lxc: linux Container library
 *
 * (C) Copyright IBM Corp. 2007, 2008
 *
 * Authors:
 * Daniel Lezcano <daniel.lezcano at free.fr>
 *
 * This library is free software; you can redistribute it and/or
 * modify it under the terms of the GNU Lesser General Public
 * License as published by the Free Software Foundation; either
 * version 2.1 of the License, or (at your option) any later version.
 *
 * This library is distributed in the hope that it will be useful,
 * but WITHOUT ANY WARRANTY; without even the implied warranty of
 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
 * Lesser General Public License for more details.
 *
 * You should have received a copy of the GNU Lesser General Public
 * License along with this library; if not, write to the Free Software
 * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
 */
#include "config.h"

#include <stdio.h>
#include <stdlib.h>
#include <errno.h>
#include <unistd.h>
#include <string.h>
#include <dirent.h>
#include <fcntl.h>
#include <grp.h>
#include <ctype.h>
#include <sys/types.h>
#include <sys/stat.h>
#include <sys/param.h>
#include <sys/inotify.h>
#include <sys/mount.h>
#include <netinet/in.h>
#include <net/if.h>

#include "error.h"
#include "commands.h"
#include "list.h"
#include "conf.h"
#include "utils.h"
#include "log.h"
#include "cgroup.h"
#include "start.h"
#include "state.h"
#include "storage.h"

#if IS_BIONIC
#include <../include/lxcmntent.h>
#else
#include <mntent.h>
#endif

struct cgroup_hierarchy;
struct cgroup_meta_data;
struct cgroup_mount_point;

/*
 * cgroup_meta_data: the metadata about the cgroup infrastructure on this
 *                   host
 */
struct cgroup_meta_data {
	ptrdiff_t ref; /* simple refcount */
	struct cgroup_hierarchy **hierarchies;
	struct cgroup_mount_point **mount_points;
	int maximum_hierarchy;
};

/*
 * cgroup_hierarchy: describes a single cgroup hierarchy
 *                   (may have multiple mount points)
 */
struct cgroup_hierarchy {
	int index;
	bool used; /* false if the hierarchy should be ignored by lxc */
	char **subsystems;
	struct cgroup_mount_point *rw_absolute_mount_point;
	struct cgroup_mount_point *ro_absolute_mount_point;
	struct cgroup_mount_point **all_mount_points;
	size_t all_mount_point_capacity;
};

/*
 * cgroup_mount_point: a mount point to where a hierarchy
 *                     is mounted to
 */
struct cgroup_mount_point {
	struct cgroup_hierarchy *hierarchy;
	char *mount_point;
	char *mount_prefix;
	bool read_only;
	bool need_cpuset_init;
};

/*
 * cgroup_process_info: describes the membership of a
 *                      process to the different cgroup
 *                      hierarchies
 *
 * Note this is the per-process info tracked by the cgfs_ops.
 * This is not used with cgmanager.
 */
struct cgroup_process_info {
	struct cgroup_process_info *next;
	struct cgroup_meta_data *meta_ref;
	struct cgroup_hierarchy *hierarchy;
	char *cgroup_path;
	char *cgroup_path_sub;
	char **created_paths;
	size_t created_paths_capacity;
	size_t created_paths_count;
	struct cgroup_mount_point *designated_mount_point;
};

struct cgfs_data {
	char *name;
	const char *cgroup_pattern;
	struct cgroup_meta_data *meta;
	struct cgroup_process_info *info;
};

lxc_log_define(lxc_cgfs, lxc);

static struct cgroup_process_info *lxc_cgroup_process_info_getx(const char *proc_pid_cgroup_str, struct cgroup_meta_data *meta);
static char **subsystems_from_mount_options(const char *mount_options, char **kernel_list);
static void lxc_cgroup_mount_point_free(struct cgroup_mount_point *mp);
static void lxc_cgroup_hierarchy_free(struct cgroup_hierarchy *h);
static bool is_valid_cgroup(const char *name);
static int create_cgroup(struct cgroup_mount_point *mp, const char *path);
static int remove_cgroup(struct cgroup_mount_point *mp, const char *path, bool recurse,
				struct lxc_conf *conf);
static char *cgroup_to_absolute_path(struct cgroup_mount_point *mp, const char *path, const char *suffix);
static struct cgroup_process_info *find_info_for_subsystem(struct cgroup_process_info *info, const char *subsystem);
static int do_cgroup_get(const char *cgroup_path, const char *sub_filename, char *value, size_t len);
static int do_cgroup_set(const char *cgroup_path, const char *sub_filename, const char *value);
static bool cgroup_devices_has_allow_or_deny(struct cgfs_data *d, char *v, bool for_allow);
static int do_setup_cgroup_limits(struct cgfs_data *d, struct lxc_list *cgroup_settings, bool do_devices);
static int cgroup_recursive_task_count(const char *cgroup_path);
static int handle_cgroup_settings(struct cgroup_mount_point *mp, char *cgroup_path);
static bool init_cpuset_if_needed(struct cgroup_mount_point *mp, const char *path);

static struct cgroup_meta_data *lxc_cgroup_load_meta2(const char **subsystem_whitelist);
static struct cgroup_meta_data *lxc_cgroup_get_meta(struct cgroup_meta_data *meta_data);
static struct cgroup_meta_data *lxc_cgroup_put_meta(struct cgroup_meta_data *meta_data);

/* free process membership information */
static void lxc_cgroup_process_info_free(struct cgroup_process_info *info);
static void lxc_cgroup_process_info_free_and_remove(struct cgroup_process_info *info,
				struct lxc_conf *conf);

static struct cgroup_ops cgfs_ops;

static int cgroup_rmdir(char *dirname)
{
	struct dirent *direntp;
	int saved_errno = 0;
	DIR *dir;
	int ret, failed=0;
	char pathname[MAXPATHLEN];

	dir = opendir(dirname);
	if (!dir) {
		ERROR("Failed to open %s", dirname);
		return -1;
	}

	while ((direntp = readdir(dir))) {
		struct stat mystat;
		int rc;

		if (!direntp)
			break;

		if (!strcmp(direntp->d_name, ".") ||
		    !strcmp(direntp->d_name, ".."))
			continue;

		rc = snprintf(pathname, MAXPATHLEN, "%s/%s", dirname, direntp->d_name);
		if (rc < 0 || rc >= MAXPATHLEN) {
			ERROR("pathname too long");
			failed=1;
			if (!saved_errno)
				saved_errno = -ENOMEM;
			continue;
		}
		ret = lstat(pathname, &mystat);
		if (ret) {
			SYSERROR("Failed to stat %s", pathname);
			failed=1;
			if (!saved_errno)
				saved_errno = errno;
			continue;
		}
		if (S_ISDIR(mystat.st_mode)) {
			if (cgroup_rmdir(pathname) < 0) {
				if (!saved_errno)
					saved_errno = errno;
				failed=1;
			}
		}
	}

	if (rmdir(dirname) < 0) {
		SYSERROR("Failed to delete %s", dirname);
		if (!saved_errno)
			saved_errno = errno;
		failed=1;
	}

	ret = closedir(dir);
	if (ret) {
		SYSERROR("Failed to close directory %s", dirname);
		if (!saved_errno)
			saved_errno = errno;
		failed=1;
	}

	errno = saved_errno;
	return failed ? -1 : 0;
}

static int rmdir_wrapper(void *data)
{
	char *path = data;

	if (setresgid(0,0,0) < 0)
		SYSERROR("Failed to setgid to 0");
	if (setresuid(0,0,0) < 0)
		SYSERROR("Failed to setuid to 0");
	if (setgroups(0, NULL) < 0)
		SYSERROR("Failed to clear groups");

	return cgroup_rmdir(path);
}

static struct cgroup_meta_data *lxc_cgroup_load_meta()
{
	const char *cgroup_use = NULL;
	char **cgroup_use_list = NULL;
	struct cgroup_meta_data *md = NULL;
	int saved_errno;

	errno = 0;
	cgroup_use = lxc_global_config_value("lxc.cgroup.use");
	if (!cgroup_use && errno != 0)
		return NULL;
	if (cgroup_use) {
		cgroup_use_list = lxc_string_split_and_trim(cgroup_use, ',');
		if (!cgroup_use_list)
			return NULL;
	}

	md = lxc_cgroup_load_meta2((const char **)cgroup_use_list);
	saved_errno = errno;
	lxc_free_array((void **)cgroup_use_list, free);
	errno = saved_errno;
	return md;
}

/* Step 1: determine all kernel subsystems */
static bool find_cgroup_subsystems(char ***kernel_subsystems)
{
	FILE *proc_cgroups;
	bool bret = false;
	char *line = NULL;
	size_t sz = 0;
	size_t kernel_subsystems_count = 0;
	size_t kernel_subsystems_capacity = 0;
	int r;

	proc_cgroups = fopen_cloexec("/proc/cgroups", "r");
	if (!proc_cgroups)
		return false;

	while (getline(&line, &sz, proc_cgroups) != -1) {
		char *tab1;
		char *tab2;
		int hierarchy_number;

		if (line[0] == '#')
			continue;
		if (!line[0])
			continue;

		tab1 = strchr(line, '\t');
		if (!tab1)
			continue;
		*tab1++ = '\0';
		tab2 = strchr(tab1, '\t');
		if (!tab2)
			continue;
		*tab2 = '\0';

		tab2 = NULL;
		hierarchy_number = strtoul(tab1, &tab2, 10);
		if (!tab2 || *tab2)
			continue;
		(void)hierarchy_number;

		r = lxc_grow_array((void ***)kernel_subsystems, &kernel_subsystems_capacity, kernel_subsystems_count + 1, 12);
		if (r < 0)
			goto out;
		(*kernel_subsystems)[kernel_subsystems_count] = strdup(line);
		if (!(*kernel_subsystems)[kernel_subsystems_count])
			goto out;
		kernel_subsystems_count++;
	}
	bret = true;

out:
	fclose(proc_cgroups);
	free(line);
	return bret;
}

/* Step 2: determine all hierarchies (by reading /proc/self/cgroup),
 *         since mount points don't specify hierarchy number and
 *         /proc/cgroups does not contain named hierarchies
 */
static bool find_cgroup_hierarchies(struct cgroup_meta_data *meta_data,
	bool all_kernel_subsystems, bool all_named_subsystems,
	const char **subsystem_whitelist)
{
	FILE *proc_self_cgroup;
	char *line = NULL;
	size_t sz = 0;
	int r;
	bool bret = false;
	size_t hierarchy_capacity = 0;

	proc_self_cgroup = fopen_cloexec("/proc/self/cgroup", "r");
	/* if for some reason (because of setns() and pid namespace for example),
	 * /proc/self is not valid, we try /proc/1/cgroup... */
	if (!proc_self_cgroup)
		proc_self_cgroup = fopen_cloexec("/proc/1/cgroup", "r");
	if (!proc_self_cgroup)
		return false;

	while (getline(&line, &sz, proc_self_cgroup) != -1) {
		/* file format: hierarchy:subsystems:group,
		 * we only extract hierarchy and subsystems
		 * here */
		char *colon1;
		char *colon2;
		int hierarchy_number;
		struct cgroup_hierarchy *h = NULL;
		char **p;

		if (!line[0])
			continue;

		colon1 = strchr(line, ':');
		if (!colon1)
			continue;
		*colon1++ = '\0';
		colon2 = strchr(colon1, ':');
		if (!colon2)
			continue;
		*colon2 = '\0';

		colon2 = NULL;

		/* With cgroupv2 /proc/self/cgroup can contain entries of the
		 * form: 0::/
		 * These entries need to be skipped.
		 */
		if (!strcmp(colon1, ""))
			continue;

		hierarchy_number = strtoul(line, &colon2, 10);
		if (!colon2 || *colon2)
			continue;

		if (hierarchy_number > meta_data->maximum_hierarchy) {
			/* lxc_grow_array will never shrink, so even if we find a lower
			* hierarchy number here, the array will never be smaller
			*/
			r = lxc_grow_array((void ***)&meta_data->hierarchies, &hierarchy_capacity, hierarchy_number + 1, 12);
			if (r < 0)
				goto out;

			meta_data->maximum_hierarchy = hierarchy_number;
		}

		/* this shouldn't happen, we had this already */
		if (meta_data->hierarchies[hierarchy_number])
			goto out;

		h = calloc(1, sizeof(struct cgroup_hierarchy));
		if (!h)
			goto out;

		meta_data->hierarchies[hierarchy_number] = h;

		h->index = hierarchy_number;
		h->subsystems = lxc_string_split_and_trim(colon1, ',');
		if (!h->subsystems)
			goto out;
		/* see if this hierarchy should be considered */
		if (!all_kernel_subsystems || !all_named_subsystems) {
			for (p = h->subsystems; *p; p++) {
				if (!strncmp(*p, "name=", 5)) {
					if (all_named_subsystems || (subsystem_whitelist && lxc_string_in_array(*p, subsystem_whitelist))) {
						h->used = true;
						break;
					}
				} else {
					if (all_kernel_subsystems || (subsystem_whitelist && lxc_string_in_array(*p, subsystem_whitelist))) {
						h->used = true;
						break;
					}
				}
			}
		} else {
			/* we want all hierarchy anyway */
			h->used = true;
		}
	}
	bret = true;

out:
	fclose(proc_self_cgroup);
	free(line);
	return bret;
}

/* Step 3: determine all mount points of each hierarchy */
static bool find_hierarchy_mountpts( struct cgroup_meta_data *meta_data, char **kernel_subsystems)
{
	bool bret = false;
	FILE *proc_self_mountinfo;
	char *line = NULL;
	size_t sz = 0;
	char **tokens = NULL;
	size_t mount_point_count = 0;
	size_t mount_point_capacity = 0;
	size_t token_capacity = 0;
	int r;
	bool is_cgns = cgns_supported();

	proc_self_mountinfo = fopen_cloexec("/proc/self/mountinfo", "r");
	/* if for some reason (because of setns() and pid namespace for example),
	 * /proc/self is not valid, we try /proc/1/cgroup... */
	if (!proc_self_mountinfo)
		proc_self_mountinfo = fopen_cloexec("/proc/1/mountinfo", "r");
	if (!proc_self_mountinfo)
		return false;

	while (getline(&line, &sz, proc_self_mountinfo) != -1) {
		char *token, *line_tok, *saveptr = NULL;
		size_t i, j, k;
		struct cgroup_mount_point *mount_point;
		struct cgroup_hierarchy *h;
		char **subsystems;
		bool is_lxcfs = false;

		if (line[0] && line[strlen(line) - 1] == '\n')
			line[strlen(line) - 1] = '\0';

		for (i = 0, line_tok = line; (token = strtok_r(line_tok, " ", &saveptr)); line_tok = NULL) {
			r = lxc_grow_array((void ***)&tokens, &token_capacity, i + 1, 64);
			if (r < 0)
				goto out;
			tokens[i++] = token;
		}

		/* layout of /proc/self/mountinfo:
		 *      0: id
		 *      1: parent id
		 *      2: device major:minor
		 *      3: mount prefix
		 *      4: mount point
		 *      5: per-mount options
		 *    [optional X]: additional data
		 *    X+7: "-"
		 *    X+8: type
		 *    X+9: source
		 *    X+10: per-superblock options
		 */
		for (j = 6; j < i && tokens[j]; j++)
			if (!strcmp(tokens[j], "-"))
				break;

		/* could not find separator */
		if (j >= i || !tokens[j])
			continue;
		/* there should be exactly three fields after
		 * the separator
		 */
		if (i != j + 4)
			continue;

		/* not a cgroup filesystem */
		if (strcmp(tokens[j + 1], "cgroup") != 0) {
			if (strcmp(tokens[j + 1], "fuse.lxcfs") != 0)
				continue;
			if (strncmp(tokens[4], "/sys/fs/cgroup/", 15) != 0)
				continue;
			is_lxcfs = true;
			char *curtok = tokens[4] + 15;
			subsystems = subsystems_from_mount_options(curtok,
							 kernel_subsystems);
		} else
			subsystems = subsystems_from_mount_options(tokens[j + 3],
							 kernel_subsystems);
		if (!subsystems)
			goto out;

		h = NULL;
		for (k = 0; k <= meta_data->maximum_hierarchy; k++) {
			if (meta_data->hierarchies[k] &&
			    meta_data->hierarchies[k]->subsystems[0] &&
			    lxc_string_in_array(meta_data->hierarchies[k]->subsystems[0], (const char **)subsystems)) {
				/* TODO: we could also check if the lists really match completely,
				 *       just to have an additional sanity check */
				h = meta_data->hierarchies[k];
				break;
			}
		}
		lxc_free_array((void **)subsystems, free);

		r = lxc_grow_array((void ***)&meta_data->mount_points, &mount_point_capacity, mount_point_count + 1, 12);
		if (r < 0)
			goto out;

		/* create mount point object */
		mount_point = calloc(1, sizeof(*mount_point));
		if (!mount_point)
			goto out;

		meta_data->mount_points[mount_point_count++] = mount_point;

		mount_point->hierarchy = h;
		if (is_lxcfs || is_cgns)
			mount_point->mount_prefix = strdup("/");
		else
			mount_point->mount_prefix = strdup(tokens[3]);
		mount_point->mount_point = strdup(tokens[4]);
		if (!mount_point->mount_point || !mount_point->mount_prefix)
			goto out;
		mount_point->read_only = !lxc_string_in_list("rw", tokens[5], ',');

		if (!strcmp(mount_point->mount_prefix, "/")) {
			if (mount_point->read_only) {
				if (!h->ro_absolute_mount_point)
					h->ro_absolute_mount_point = mount_point;
			} else {
				if (!h->rw_absolute_mount_point)
					h->rw_absolute_mount_point = mount_point;
			}
		}

		if (h)
			k = lxc_array_len((void **)h->all_mount_points);
		else
			k = 0;
		r = lxc_grow_array((void ***)&h->all_mount_points, &h->all_mount_point_capacity, k + 1, 4);
		if (r < 0)
			goto out;
		h->all_mount_points[k] = mount_point;
	}
	bret = true;

out:
	fclose(proc_self_mountinfo);
	free(tokens);
	free(line);
	return bret;
}

static struct cgroup_meta_data *lxc_cgroup_load_meta2(const char **subsystem_whitelist)
{
	bool all_kernel_subsystems = true;
	bool all_named_subsystems = false;
	struct cgroup_meta_data *meta_data = NULL;
	char **kernel_subsystems = NULL;
	int saved_errno = 0;

	/* if the subsystem whitelist is not specified, include all
	 * hierarchies that contain kernel subsystems by default but
	 * no hierarchies that only contain named subsystems
	 *
	 * if it is specified, the specifier @all will select all
	 * hierarchies, @kernel will select all hierarchies with
	 * kernel subsystems and @named will select all named
	 * hierarchies
	 */
	all_kernel_subsystems = subsystem_whitelist ?
		(lxc_string_in_array("@kernel", subsystem_whitelist) || lxc_string_in_array("@all", subsystem_whitelist)) :
		true;
	all_named_subsystems = subsystem_whitelist ?
		(lxc_string_in_array("@named", subsystem_whitelist) || lxc_string_in_array("@all", subsystem_whitelist)) :
		true;

	meta_data = calloc(1, sizeof(struct cgroup_meta_data));
	if (!meta_data)
		return NULL;
	meta_data->ref = 1;

	if (!find_cgroup_subsystems(&kernel_subsystems))
		goto out_error;

	if (!find_cgroup_hierarchies(meta_data, all_kernel_subsystems,
				all_named_subsystems, subsystem_whitelist))
		goto out_error;

	if (!find_hierarchy_mountpts(meta_data, kernel_subsystems))
		goto out_error;

	/* oops, we couldn't find anything */
	if (!meta_data->hierarchies || !meta_data->mount_points) {
		errno = EINVAL;
		goto out_error;
	}

	lxc_free_array((void **)kernel_subsystems, free);
	return meta_data;

out_error:
	saved_errno = errno;
	lxc_free_array((void **)kernel_subsystems, free);
	lxc_cgroup_put_meta(meta_data);
	errno = saved_errno;
	return NULL;
}

static struct cgroup_meta_data *lxc_cgroup_get_meta(struct cgroup_meta_data *meta_data)
{
	meta_data->ref++;
	return meta_data;
}

static struct cgroup_meta_data *lxc_cgroup_put_meta(struct cgroup_meta_data *meta_data)
{
	size_t i;
	if (!meta_data)
		return NULL;
	if (--meta_data->ref > 0)
		return meta_data;
	lxc_free_array((void **)meta_data->mount_points, (lxc_free_fn)lxc_cgroup_mount_point_free);
	if (meta_data->hierarchies)
		for (i = 0; i <= meta_data->maximum_hierarchy; i++)
			if (meta_data->hierarchies[i])
				lxc_cgroup_hierarchy_free(meta_data->hierarchies[i]);
	free(meta_data->hierarchies);
	free(meta_data);
	return NULL;
}

static struct cgroup_hierarchy *lxc_cgroup_find_hierarchy(struct cgroup_meta_data *meta_data, const char *subsystem)
{
	size_t i;
	for (i = 0; i <= meta_data->maximum_hierarchy; i++) {
		struct cgroup_hierarchy *h = meta_data->hierarchies[i];
		if (!h)
			continue;
		if (h && lxc_string_in_array(subsystem, (const char **)h->subsystems))
			return h;
	}
	return NULL;
}

static bool mountpoint_is_accessible(struct cgroup_mount_point *mp)
{
	return mp && access(mp->mount_point, F_OK) == 0;
}

static struct cgroup_mount_point *lxc_cgroup_find_mount_point(struct cgroup_hierarchy *hierarchy, const char *group, bool should_be_writable)
{
	struct cgroup_mount_point **mps;
	struct cgroup_mount_point *current_result = NULL;
	ssize_t quality = -1;

	/* trivial case */
	if (mountpoint_is_accessible(hierarchy->rw_absolute_mount_point))
		return hierarchy->rw_absolute_mount_point;
	if (!should_be_writable && mountpoint_is_accessible(hierarchy->ro_absolute_mount_point))
		return hierarchy->ro_absolute_mount_point;

	for (mps = hierarchy->all_mount_points; mps && *mps; mps++) {
		struct cgroup_mount_point *mp = *mps;
		size_t prefix_len = mp->mount_prefix ? strlen(mp->mount_prefix) : 0;

		if (prefix_len == 1 && mp->mount_prefix[0] == '/')
			prefix_len = 0;

		if (!mountpoint_is_accessible(mp))
			continue;

		if (should_be_writable && mp->read_only)
			continue;

		if (!prefix_len ||
		    (strncmp(group, mp->mount_prefix, prefix_len) == 0 &&
		     (group[prefix_len] == '\0' || group[prefix_len] == '/'))) {
			/* search for the best quality match, i.e. the match with the
			 * shortest prefix where this group is still contained
			 */
			if (quality == -1 || prefix_len < quality) {
				current_result = mp;
				quality = prefix_len;
			}
		}
	}

	if (!current_result)
		errno = ENOENT;
	return current_result;
}

static char *lxc_cgroup_find_abs_path(const char *subsystem, const char *group, bool should_be_writable, const char *suffix)
{
	struct cgroup_meta_data *meta_data;
	struct cgroup_hierarchy *h;
	struct cgroup_mount_point *mp;
	char *result;
	int saved_errno;

	meta_data = lxc_cgroup_load_meta();
	if (!meta_data)
		return NULL;

	h = lxc_cgroup_find_hierarchy(meta_data, subsystem);
	if (!h)
		goto out_error;

	mp = lxc_cgroup_find_mount_point(h, group, should_be_writable);
	if (!mp)
		goto out_error;

	result = cgroup_to_absolute_path(mp, group, suffix);
	if (!result)
		goto out_error;

	lxc_cgroup_put_meta(meta_data);
	return result;

out_error:
	saved_errno = errno;
	lxc_cgroup_put_meta(meta_data);
	errno = saved_errno;
	return NULL;
}

static struct cgroup_process_info *lxc_cgroup_process_info_get(pid_t pid, struct cgroup_meta_data *meta)
{
	char pid_buf[32];
	snprintf(pid_buf, 32, "/proc/%lu/cgroup", (unsigned long)pid);
	return lxc_cgroup_process_info_getx(pid_buf, meta);
}

static struct cgroup_process_info *lxc_cgroup_process_info_get_init(struct cgroup_meta_data *meta)
{
	return lxc_cgroup_process_info_get(1, meta);
}

static struct cgroup_process_info *lxc_cgroup_process_info_get_self(struct cgroup_meta_data *meta)
{
	struct cgroup_process_info *i;
	i = lxc_cgroup_process_info_getx("/proc/self/cgroup", meta);
	if (!i)
		i = lxc_cgroup_process_info_get(lxc_raw_getpid(), meta);
	return i;
}

/*
 * If a controller has ns cgroup mounted, then in that cgroup the handler->pid
 * is already in a new cgroup named after the pid.  'mnt' is passed in as
 * the full current cgroup.  Say that is /sys/fs/cgroup/lxc/2975 and the container
 * name is c1. .  We want to rename the cgroup directory to /sys/fs/cgroup/lxc/c1,
 * and return the string /sys/fs/cgroup/lxc/c1.
 */
static char *cgroup_rename_nsgroup(const char *mountpath, const char *oldname, pid_t pid, const char *name)
{
	char *dir, *fulloldpath;
	char *newname, *fullnewpath;
	int len, newlen, ret;

	/*
	 * if cgroup is mounted at /cgroup and task is in cgroup /ab/, pid 2375 and
	 * name is c1,
	 * dir: /ab
	 * fulloldpath = /cgroup/ab/2375
	 * fullnewpath = /cgroup/ab/c1
	 * newname = /ab/c1
	 */
	dir = alloca(strlen(oldname) + 1);
	strcpy(dir, oldname);

	len = strlen(oldname) + strlen(mountpath) + 22;
	fulloldpath = alloca(len);
	ret = snprintf(fulloldpath, len, "%s/%s/%lu", mountpath, oldname, (unsigned long)pid);
	if (ret < 0 || ret >= len)
		return NULL;

	len = strlen(dir) + strlen(name) + 2;
	newname = malloc(len);
	if (!newname) {
		SYSERROR("Out of memory");
		return NULL;
	}
	ret = snprintf(newname, len, "%s/%s", dir, name);
	if (ret < 0 || ret >= len) {
		free(newname);
		return NULL;
	}

	newlen = strlen(mountpath) + len + 2;
	fullnewpath = alloca(newlen);
	ret = snprintf(fullnewpath, newlen, "%s/%s", mountpath, newname);
	if (ret < 0 || ret >= newlen) {
		free(newname);
		return NULL;
	}

	if (access(fullnewpath, F_OK) == 0) {
		if (rmdir(fullnewpath) != 0) {
			SYSERROR("container cgroup %s already exists.", fullnewpath);
			free(newname);
			return NULL;
		}
	}
	if (rename(fulloldpath, fullnewpath)) {
		SYSERROR("failed to rename cgroup %s->%s", fulloldpath, fullnewpath);
		free(newname);
		return NULL;
	}

	DEBUG("'%s' renamed to '%s'", oldname, newname);

	return newname;
}

static bool is_crucial_hierarchy(struct cgroup_hierarchy *h)
{
	char **p;

	for (p = h->subsystems; *p; p++) {
		if (is_crucial_cgroup_subsystem(*p))
			return true;
	}
	return false;
}

/* create a new cgroup */
static struct cgroup_process_info *lxc_cgroupfs_create(const char *name, const char *path_pattern, struct cgroup_meta_data *meta_data, const char *sub_pattern)
{
	char **cgroup_path_components = NULL;
	char **p = NULL;
	char *path_so_far = NULL;
	char **new_cgroup_paths = NULL;
	char **new_cgroup_paths_sub = NULL;
	struct cgroup_mount_point *mp;
	struct cgroup_hierarchy *h;
	struct cgroup_process_info *base_info = NULL;
	struct cgroup_process_info *info_ptr;
	int saved_errno;
	int r;
	unsigned suffix = 0;
	bool had_sub_pattern = false;
	size_t i;

	if (!is_valid_cgroup(name)) {
		ERROR("Invalid cgroup name: '%s'", name);
		errno = EINVAL;
		return NULL;
	}

	if (!strstr(path_pattern, "%n")) {
		ERROR("Invalid cgroup path pattern: '%s'; contains no %%n for specifying container name", path_pattern);
		errno = EINVAL;
		return NULL;
	}

	/* we will modify the result of this operation directly,
	 * so we don't have to copy the data structure
	 */
	base_info = (path_pattern[0] == '/') ?
		lxc_cgroup_process_info_get_init(meta_data) :
		lxc_cgroup_process_info_get_self(meta_data);
	if (!base_info)
		return NULL;

	new_cgroup_paths = calloc(meta_data->maximum_hierarchy + 1, sizeof(char *));
	if (!new_cgroup_paths)
		goto out_initial_error;

	new_cgroup_paths_sub = calloc(meta_data->maximum_hierarchy + 1, sizeof(char *));
	if (!new_cgroup_paths_sub)
		goto out_initial_error;

	/* find mount points we can use */
	for (info_ptr = base_info; info_ptr; info_ptr = info_ptr->next) {
		h = info_ptr->hierarchy;
		if (!h)
			continue;
		mp = lxc_cgroup_find_mount_point(h, info_ptr->cgroup_path, true);
		if (!mp) {
			ERROR("Could not find writable mount point for cgroup hierarchy %d while trying to create cgroup.", h->index);
			goto out_initial_error;
		}
		info_ptr->designated_mount_point = mp;

		if (lxc_string_in_array("ns", (const char **)h->subsystems))
			continue;
		if (handle_cgroup_settings(mp, info_ptr->cgroup_path) < 0) {
			ERROR("Could not set clone_children to 1 for cpuset hierarchy in parent cgroup.");
			goto out_initial_error;
		}
	}

	/* normalize the path */
	cgroup_path_components = lxc_normalize_path(path_pattern);
	if (!cgroup_path_components)
		goto out_initial_error;

	/* go through the path components to see if we can create them */
	for (p = cgroup_path_components; *p || (sub_pattern && !had_sub_pattern); p++) {
		/* we only want to create the same component with -1, -2, etc.
		 * if the component contains the container name itself, otherwise
		 * it's not an error if it already exists
		 */
		char *p_eff = *p ? *p : (char *)sub_pattern;
		bool contains_name = strstr(p_eff, "%n");
		char *current_component = NULL;
		char *current_subpath = NULL;
		char *current_entire_path = NULL;
		char *parts[3];
		size_t j = 0;
		i = 0;

		/* if we are processing the subpattern, we want to make sure
		 * loop is ended the next time around
		 */
		if (!*p) {
			had_sub_pattern = true;
			p--;
		}

		goto find_name_on_this_level;

	cleanup_name_on_this_level:
		/* This is reached if we found a name clash.
		 * In that case, remove the cgroup from all previous hierarchies
		 */
		for (j = 0, info_ptr = base_info; j < i && info_ptr; info_ptr = info_ptr->next, j++) {
			if (info_ptr->created_paths_count < 1)
				continue;
			r = remove_cgroup(info_ptr->designated_mount_point, info_ptr->created_paths[info_ptr->created_paths_count - 1], false, NULL);
			if (r < 0)
				WARN("could not clean up cgroup we created when trying to create container");
			free(info_ptr->created_paths[info_ptr->created_paths_count - 1]);
			info_ptr->created_paths[--info_ptr->created_paths_count] = NULL;
		}
		if (current_component != current_subpath)
			free(current_subpath);
		if (current_component != p_eff)
			free(current_component);
		current_component = current_subpath = NULL;
		/* try again with another suffix */
		++suffix;

	find_name_on_this_level:
		/* determine name of the path component we should create */
		if (contains_name && suffix > 0) {
			char *buf = calloc(strlen(name) + 32, 1);
			if (!buf)
				goto out_initial_error;
			snprintf(buf, strlen(name) + 32, "%s-%u", name, suffix);
			current_component = lxc_string_replace("%n", buf, p_eff);
			free(buf);
		} else {
			current_component = contains_name ? lxc_string_replace("%n", name, p_eff) : p_eff;
		}
		parts[0] = path_so_far;
		parts[1] = current_component;
		parts[2] = NULL;
		current_subpath = path_so_far ? lxc_string_join("/", (const char **)parts, false) : current_component;

		/* Now go through each hierarchy and try to create the
		 * corresponding cgroup
		 */
		for (i = 0, info_ptr = base_info; info_ptr; info_ptr = info_ptr->next, i++) {
			char *parts2[3];

			if (!info_ptr->hierarchy)
				continue;

			if (lxc_string_in_array("ns", (const char **)info_ptr->hierarchy->subsystems))
				continue;
			current_entire_path = NULL;

			parts2[0] = !strcmp(info_ptr->cgroup_path, "/") ? "" : info_ptr->cgroup_path;
			parts2[1] = current_subpath;
			parts2[2] = NULL;
			current_entire_path = lxc_string_join("/", (const char **)parts2, false);

			if (!*p) {
				/* we are processing the subpath, so only update that one */
				free(new_cgroup_paths_sub[i]);
				new_cgroup_paths_sub[i] = strdup(current_entire_path);
				if (!new_cgroup_paths_sub[i])
					goto cleanup_from_error;
			} else {
				/* remember which path was used on this controller */
				free(new_cgroup_paths[i]);
				new_cgroup_paths[i] = strdup(current_entire_path);
				if (!new_cgroup_paths[i])
					goto cleanup_from_error;
			}

			r = create_cgroup(info_ptr->designated_mount_point, current_entire_path);
			if (r < 0 && errno == EEXIST && contains_name) {
				/* name clash => try new name with new suffix */
				free(current_entire_path);
				current_entire_path = NULL;
				goto cleanup_name_on_this_level;
			} else if (r < 0 && errno != EEXIST) {
				if (is_crucial_hierarchy(info_ptr->hierarchy)) {
					SYSERROR("Could not create cgroup '%s' in '%s'.", current_entire_path, info_ptr->designated_mount_point->mount_point);
					goto cleanup_from_error;
				}
				goto skip;
			} else if (r == 0) {
				/* successfully created */
				r = lxc_grow_array((void ***)&info_ptr->created_paths, &info_ptr->created_paths_capacity, info_ptr->created_paths_count + 1, 8);
				if (r < 0)
					goto cleanup_from_error;
				if (!init_cpuset_if_needed(info_ptr->designated_mount_point, current_entire_path)) {
					ERROR("Failed to initialize cpuset for '%s' in '%s'.", current_entire_path, info_ptr->designated_mount_point->mount_point);
					goto cleanup_from_error;
				}
				info_ptr->created_paths[info_ptr->created_paths_count++] = current_entire_path;
			} else {
				/* if we didn't create the cgroup, then we have to make sure that
				 * further cgroups will be created properly
				 */
				if (handle_cgroup_settings(info_ptr->designated_mount_point, info_ptr->cgroup_path) < 0) {
					ERROR("Could not set clone_children to 1 for cpuset hierarchy in pre-existing cgroup.");
					goto cleanup_from_error;
				}
				if (!init_cpuset_if_needed(info_ptr->designated_mount_point, info_ptr->cgroup_path)) {
					ERROR("Failed to initialize cpuset in pre-existing '%s'.", info_ptr->cgroup_path);
					goto cleanup_from_error;
				}

skip:
				/* already existed but path component of pattern didn't contain '%n',
				 * so this is not an error; but then we don't need current_entire_path
				 * anymore...
				 */
				free(current_entire_path);
				current_entire_path = NULL;
			}
		}

		/* save path so far */
		free(path_so_far);
		path_so_far = strdup(current_subpath);
		if (!path_so_far)
			goto cleanup_from_error;

		/* cleanup */
		if (current_component != current_subpath)
			free(current_subpath);
		if (current_component != p_eff)
			free(current_component);
		current_component = current_subpath = NULL;
		continue;

	cleanup_from_error:
		/* called if an error occurred in the loop, so we
		 * do some additional cleanup here
		 */
		saved_errno = errno;
		if (current_component != current_subpath)
			free(current_subpath);
		if (current_component != p_eff)
			free(current_component);
		free(current_entire_path);
		errno = saved_errno;
		goto out_initial_error;
	}

	/* we're done, now update the paths */
	for (i = 0, info_ptr = base_info; info_ptr; info_ptr = info_ptr->next, i++) {
		if (!info_ptr->hierarchy)
			continue;
		/* ignore legacy 'ns' subsystem here, lxc_cgroup_create_legacy
		 * will take care of it
		 * Since we do a continue in above loop, new_cgroup_paths[i] is
		 * unset anyway, as is new_cgroup_paths_sub[i]
		 */
		if (lxc_string_in_array("ns", (const char **)info_ptr->hierarchy->subsystems))
			continue;
		free(info_ptr->cgroup_path);
		info_ptr->cgroup_path = new_cgroup_paths[i];
		info_ptr->cgroup_path_sub = new_cgroup_paths_sub[i];
	}
	/* don't use lxc_free_array since we used the array members
	 * to store them in our result...
	 */
	free(new_cgroup_paths);
	free(new_cgroup_paths_sub);
	free(path_so_far);
	lxc_free_array((void **)cgroup_path_components, free);
	return base_info;

out_initial_error:
	saved_errno = errno;
	free(path_so_far);
	lxc_cgroup_process_info_free_and_remove(base_info, NULL);
	lxc_free_array((void **)new_cgroup_paths, free);
	lxc_free_array((void **)new_cgroup_paths_sub, free);
	lxc_free_array((void **)cgroup_path_components, free);
	errno = saved_errno;
	return NULL;
}

static int lxc_cgroup_create_legacy(struct cgroup_process_info *base_info, const char *name, pid_t pid)
{
	struct cgroup_process_info *info_ptr;
	int r;

	for (info_ptr = base_info; info_ptr; info_ptr = info_ptr->next) {
		if (!info_ptr->hierarchy)
			continue;

		if (!lxc_string_in_array("ns", (const char **)info_ptr->hierarchy->subsystems))
			continue;
		/*
		 * For any path which has ns cgroup mounted, handler->pid is already
		 * moved into a container called '%d % (handler->pid)'.  Rename it to
		 * the cgroup name and record that.
		 */
		char *tmp = cgroup_rename_nsgroup((const char *)info_ptr->designated_mount_point->mount_point,
				info_ptr->cgroup_path, pid, name);
		if (!tmp)
			return -1;
		free(info_ptr->cgroup_path);
		info_ptr->cgroup_path = tmp;
		r = lxc_grow_array((void ***)&info_ptr->created_paths, &info_ptr->created_paths_capacity, info_ptr->created_paths_count + 1, 8);
		if (r < 0)
			return -1;
		tmp = strdup(tmp);
		if (!tmp)
			return -1;
		info_ptr->created_paths[info_ptr->created_paths_count++] = tmp;
	}
	return 0;
}

/* get the cgroup membership of a given container */
static struct cgroup_process_info *lxc_cgroup_get_container_info(const char *name, const char *lxcpath, struct cgroup_meta_data *meta_data)
{
	struct cgroup_process_info *result = NULL;
	int saved_errno = 0;
	size_t i;
	struct cgroup_process_info **cptr = &result;
	struct cgroup_process_info *entry = NULL;
	char *path = NULL;

	for (i = 0; i <= meta_data->maximum_hierarchy; i++) {
		struct cgroup_hierarchy *h = meta_data->hierarchies[i];
		if (!h || !h->used)
			continue;

		/* use the command interface to look for the cgroup */
		path = lxc_cmd_get_cgroup_path(name, lxcpath, h->subsystems[0]);
		if (!path) {
			h->used = false;
			continue;
		}

		entry = calloc(1, sizeof(struct cgroup_process_info));
		if (!entry)
			goto out_error;
		entry->meta_ref = lxc_cgroup_get_meta(meta_data);
		entry->hierarchy = h;
		entry->cgroup_path = path;
		path = NULL;

		/* it is not an error if we don't find anything here,
		 * it is up to the caller to decide what to do in that
		 * case */
		entry->designated_mount_point = lxc_cgroup_find_mount_point(h, entry->cgroup_path, true);

		*cptr = entry;
		cptr = &entry->next;
		entry = NULL;
	}

	return result;
out_error:
	saved_errno = errno;
	free(path);
	lxc_cgroup_process_info_free(result);
	lxc_cgroup_process_info_free(entry);
	errno = saved_errno;
	return NULL;
}

/* move a processs to the cgroups specified by the membership */
static int lxc_cgroupfs_enter(struct cgroup_process_info *info, pid_t pid, bool enter_sub)
{
	char pid_buf[32];
	char *cgroup_tasks_fn;
	int r;
	struct cgroup_process_info *info_ptr;

	snprintf(pid_buf, 32, "%lu", (unsigned long)pid);
	for (info_ptr = info; info_ptr; info_ptr = info_ptr->next) {
		if (!info_ptr->hierarchy)
			continue;

		char *cgroup_path = (enter_sub && info_ptr->cgroup_path_sub) ?
			info_ptr->cgroup_path_sub :
			info_ptr->cgroup_path;

		if (!info_ptr->designated_mount_point) {
			info_ptr->designated_mount_point = lxc_cgroup_find_mount_point(info_ptr->hierarchy, cgroup_path, true);
			if (!info_ptr->designated_mount_point) {
				SYSERROR("Could not add pid %lu to cgroup %s: internal error (couldn't find any writable mountpoint to cgroup filesystem)", (unsigned long)pid, cgroup_path);
				return -1;
			}
		}

		cgroup_tasks_fn = cgroup_to_absolute_path(info_ptr->designated_mount_point, cgroup_path, "/tasks");
		if (!cgroup_tasks_fn) {
			SYSERROR("Could not add pid %lu to cgroup %s: internal error", (unsigned long)pid, cgroup_path);
			return -1;
		}

		r = lxc_write_to_file(cgroup_tasks_fn, pid_buf, strlen(pid_buf), false);
		free(cgroup_tasks_fn);
		if (r < 0 && is_crucial_hierarchy(info_ptr->hierarchy)) {
			SYSERROR("Could not add pid %lu to cgroup %s: internal error", (unsigned long)pid, cgroup_path);
			return -1;
		}
	}

	return 0;
}

/* free process membership information */
void lxc_cgroup_process_info_free(struct cgroup_process_info *info)
{
	struct cgroup_process_info *next;
	if (!info)
		return;
	next = info->next;
	lxc_cgroup_put_meta(info->meta_ref);
	free(info->cgroup_path);
	free(info->cgroup_path_sub);
	lxc_free_array((void **)info->created_paths, free);
	free(info);
	lxc_cgroup_process_info_free(next);
}

/* free process membership information and remove cgroups that were created */
void lxc_cgroup_process_info_free_and_remove(struct cgroup_process_info *info, struct lxc_conf *conf)
{
	struct cgroup_process_info *next;
	char **pp;
	if (!info)
		return;
	next = info->next;
	{
		struct cgroup_mount_point *mp = info->designated_mount_point;
		if (!mp)
			mp = lxc_cgroup_find_mount_point(info->hierarchy, info->cgroup_path, true);
		if (mp)
			/* ignore return value here, perhaps we created the
			 * '/lxc' cgroup in this container but another container
			 * is still running (for example)
			 */
			(void)remove_cgroup(mp, info->cgroup_path, true, conf);
	}
	for (pp = info->created_paths; pp && *pp; pp++);
	for ((void)(pp && --pp); info->created_paths && pp >= info->created_paths; --pp) {
		free(*pp);
	}
	free(info->created_paths);
	lxc_cgroup_put_meta(info->meta_ref);
	free(info->cgroup_path);
	free(info->cgroup_path_sub);
	free(info);
	lxc_cgroup_process_info_free_and_remove(next, conf);
}

static char *lxc_cgroup_get_hierarchy_path_data(const char *subsystem, struct cgfs_data *d)
{
	struct cgroup_process_info *info = d->info;
	info = find_info_for_subsystem(info, subsystem);
	if (!info)
		return NULL;
	prune_init_scope(info->cgroup_path);
	return info->cgroup_path;
}

static char *lxc_cgroup_get_hierarchy_abs_path_data(const char *subsystem, struct cgfs_data *d)
{
	struct cgroup_process_info *info = d->info;
	struct cgroup_mount_point *mp = NULL;

	info = find_info_for_subsystem(info, subsystem);
	if (!info)
		return NULL;
	if (info->designated_mount_point) {
		mp = info->designated_mount_point;
	} else {
		mp = lxc_cgroup_find_mount_point(info->hierarchy, info->cgroup_path, true);
		if (!mp)
			return NULL;
	}
	return cgroup_to_absolute_path(mp, info->cgroup_path, NULL);
}

static char *lxc_cgroup_get_hierarchy_abs_path(const char *subsystem, const char *name, const char *lxcpath)
{
	struct cgroup_meta_data *meta;
	struct cgroup_process_info *base_info, *info;
	struct cgroup_mount_point *mp;
	char *result = NULL;

	meta = lxc_cgroup_load_meta();
	if (!meta)
		return NULL;
	base_info = lxc_cgroup_get_container_info(name, lxcpath, meta);
	if (!base_info)
		goto out1;
	info = find_info_for_subsystem(base_info, subsystem);
	if (!info)
		goto out2;
	if (info->designated_mount_point) {
		mp = info->designated_mount_point;
	} else {
		mp = lxc_cgroup_find_mount_point(info->hierarchy, info->cgroup_path, true);
		if (!mp)
			goto out3;
	}
	result = cgroup_to_absolute_path(mp, info->cgroup_path, NULL);
out3:
out2:
	lxc_cgroup_process_info_free(base_info);
out1:
	lxc_cgroup_put_meta(meta);
	return result;
}

static int lxc_cgroup_set_data(const char *filename, const char *value, struct cgfs_data *d)
{
	char *subsystem = NULL, *p, *path;
	int ret = -1;

	subsystem = alloca(strlen(filename) + 1);
	strcpy(subsystem, filename);
	if ((p = strchr(subsystem, '.')) != NULL)
		*p = '\0';

	errno = ENOENT;
	path = lxc_cgroup_get_hierarchy_abs_path_data(subsystem, d);
	if (path) {
		ret = do_cgroup_set(path, filename, value);
		int saved_errno = errno;
		free(path);
		errno = saved_errno;
	}
	return ret;
}

static int lxc_cgroupfs_set(const char *filename, const char *value, const char *name, const char *lxcpath)
{
	char *subsystem = NULL, *p, *path;
	int ret = -1;

	subsystem = alloca(strlen(filename) + 1);
	strcpy(subsystem, filename);
	if ((p = strchr(subsystem, '.')) != NULL)
		*p = '\0';

	path = lxc_cgroup_get_hierarchy_abs_path(subsystem, name, lxcpath);
	if (path) {
		ret = do_cgroup_set(path, filename, value);
		free(path);
	}
	return ret;
}

static int lxc_cgroupfs_get(const char *filename, char *value, size_t len, const char *name, const char *lxcpath)
{
	char *subsystem = NULL, *p, *path;
	int ret = -1;

	subsystem = alloca(strlen(filename) + 1);
	strcpy(subsystem, filename);
	if ((p = strchr(subsystem, '.')) != NULL)
		*p = '\0';

	path = lxc_cgroup_get_hierarchy_abs_path(subsystem, name, lxcpath);
	if (path) {
		ret = do_cgroup_get(path, filename, value, len);
		free(path);
	}
	return ret;
}

static bool cgroupfs_mount_cgroup(void *hdata, const char *root, int type)
{
	size_t bufsz = strlen(root) + sizeof("/sys/fs/cgroup");
	char *path = NULL;
	char **parts = NULL;
	char *dirname = NULL;
	char *abs_path = NULL;
	char *abs_path2 = NULL;
	struct cgfs_data *cgfs_d;
	struct cgroup_process_info *info, *base_info;
	int r, saved_errno = 0;
	struct lxc_handler *handler = hdata;

	if (cgns_supported())
		return true;

	cgfs_d = handler->cgroup_data;
	if (!cgfs_d)
		return false;
	base_info = cgfs_d->info;

	/* If we get passed the _NOSPEC types, we default to _MIXED, since we don't
	 * have access to the lxc_conf object at this point. It really should be up
	 * to the caller to fix this, but this doesn't really hurt.
	 */
	if (type == LXC_AUTO_CGROUP_FULL_NOSPEC)
		type = LXC_AUTO_CGROUP_FULL_MIXED;
	else if (type == LXC_AUTO_CGROUP_NOSPEC)
		type = LXC_AUTO_CGROUP_MIXED;

	if (type < LXC_AUTO_CGROUP_RO || type > LXC_AUTO_CGROUP_FULL_MIXED) {
		ERROR("could not mount cgroups into container: invalid type specified internally");
		errno = EINVAL;
		return false;
	}

	path = calloc(1, bufsz);
	if (!path)
		return false;
	snprintf(path, bufsz, "%s/sys/fs/cgroup", root);
	r = safe_mount("cgroup_root", path, "tmpfs",
			MS_NOSUID|MS_NODEV|MS_NOEXEC|MS_RELATIME,
			"size=10240k,mode=755",
			root);
	if (r < 0) {
		SYSERROR("could not mount tmpfs to /sys/fs/cgroup in the container");
		return false;
	}

	/* now mount all the hierarchies we care about */
	for (info = base_info; info; info = info->next) {
		size_t subsystem_count, i;
		struct cgroup_mount_point *mp = info->designated_mount_point;

		if (!info->hierarchy)
			continue;

		if (!mountpoint_is_accessible(mp))
			mp = lxc_cgroup_find_mount_point(info->hierarchy, info->cgroup_path, true);

		if (!mp) {
			SYSERROR("could not find original mount point for cgroup hierarchy while trying to mount cgroup filesystem");
			goto out_error;
		}

		subsystem_count = lxc_array_len((void **)info->hierarchy->subsystems);
		parts = calloc(subsystem_count + 1, sizeof(char *));
		if (!parts)
			goto out_error;

		for (i = 0; i < subsystem_count; i++) {
			if (!strncmp(info->hierarchy->subsystems[i], "name=", 5))
				parts[i] = info->hierarchy->subsystems[i] + 5;
			else
				parts[i] = info->hierarchy->subsystems[i];
		}
		dirname = lxc_string_join(",", (const char **)parts, false);
		if (!dirname)
			goto out_error;

		/* create subsystem directory */
		abs_path = lxc_append_paths(path, dirname);
		if (!abs_path)
			goto out_error;
		r = mkdir_p(abs_path, 0755);
		if (r < 0 && errno != EEXIST) {
			SYSERROR("could not create cgroup subsystem directory /sys/fs/cgroup/%s", dirname);
			goto out_error;
		}

		abs_path2 = lxc_append_paths(abs_path, info->cgroup_path);
		if (!abs_path2)
			goto out_error;

		if (type == LXC_AUTO_CGROUP_FULL_RO || type == LXC_AUTO_CGROUP_FULL_RW || type == LXC_AUTO_CGROUP_FULL_MIXED) {
			/* bind-mount the cgroup entire filesystem there */
			if (strcmp(mp->mount_prefix, "/") != 0) {
				/* FIXME: maybe we should just try to remount the entire hierarchy
				 *        with a regular mount command? may that works? */
				ERROR("could not automatically mount cgroup-full to /sys/fs/cgroup/%s: host has no mount point for this cgroup filesystem that has access to the root cgroup", dirname);
				goto out_error;
			}
			r = mount(mp->mount_point, abs_path, "none", MS_BIND, 0);
			if (r < 0) {
				SYSERROR("error bind-mounting %s to %s", mp->mount_point, abs_path);
				goto out_error;
			}
			/* main cgroup path should be read-only */
			if (type == LXC_AUTO_CGROUP_FULL_RO || type == LXC_AUTO_CGROUP_FULL_MIXED) {
				r = mount(NULL, abs_path, NULL, MS_REMOUNT|MS_BIND|MS_RDONLY, NULL);
				if (r < 0) {
					SYSERROR("error re-mounting %s readonly", abs_path);
					goto out_error;
				}
			}
			/* own cgroup should be read-write */
			if (type == LXC_AUTO_CGROUP_FULL_MIXED) {
				r = mount(abs_path2, abs_path2, NULL, MS_BIND, NULL);
				if (r < 0) {
					SYSERROR("error bind-mounting %s onto itself", abs_path2);
					goto out_error;
				}
				r = mount(NULL, abs_path2, NULL, MS_REMOUNT|MS_BIND, NULL);
				if (r < 0) {
					SYSERROR("error re-mounting %s readwrite", abs_path2);
					goto out_error;
				}
			}
		} else {
			/* create path for container's cgroup */
			r = mkdir_p(abs_path2, 0755);
			if (r < 0 && errno != EEXIST) {
				SYSERROR("could not create cgroup directory /sys/fs/cgroup/%s%s", dirname, info->cgroup_path);
				goto out_error;
			}

			/* for read-only and mixed cases, we have to bind-mount the tmpfs directory
			 * that points to the hierarchy itself (i.e. /sys/fs/cgroup/cpu etc.) onto
			 * itself and then bind-mount it read-only, since we keep the tmpfs itself
			 * read-write (see comment below)
			 */
			if (type == LXC_AUTO_CGROUP_MIXED || type == LXC_AUTO_CGROUP_RO) {
				r = mount(abs_path, abs_path, NULL, MS_BIND, NULL);
				if (r < 0) {
					SYSERROR("error bind-mounting %s onto itself", abs_path);
					goto out_error;
				}
				r = mount(NULL, abs_path, NULL, MS_REMOUNT|MS_BIND|MS_RDONLY, NULL);
				if (r < 0) {
					SYSERROR("error re-mounting %s readonly", abs_path);
					goto out_error;
				}
			}

			free(abs_path);
			abs_path = NULL;

			/* bind-mount container's cgroup to that directory */
			abs_path = cgroup_to_absolute_path(mp, info->cgroup_path, NULL);
			if (!abs_path)
				goto out_error;
			r = mount(abs_path, abs_path2, "none", MS_BIND, 0);
			if (r < 0 && is_crucial_hierarchy(info->hierarchy)) {
				SYSERROR("error bind-mounting %s to %s", abs_path, abs_path2);
				goto out_error;
			}
			if (type == LXC_AUTO_CGROUP_RO) {
				r = mount(NULL, abs_path2, NULL, MS_REMOUNT|MS_BIND|MS_RDONLY, NULL);
				if (r < 0) {
					SYSERROR("error re-mounting %s readonly", abs_path2);
					goto out_error;
				}
			}
		}

		free(abs_path);
		free(abs_path2);
		abs_path = NULL;
		abs_path2 = NULL;

		/* add symlinks for every single subsystem */
		if (subsystem_count > 1) {
			for (i = 0; i < subsystem_count; i++) {
				abs_path = lxc_append_paths(path, parts[i]);
				if (!abs_path)
					goto out_error;
				r = symlink(dirname, abs_path);
				if (r < 0)
					WARN("could not create symlink %s -> %s in /sys/fs/cgroup of container", parts[i], dirname);
				free(abs_path);
				abs_path = NULL;
			}
		}
		free(dirname);
		free(parts);
		dirname = NULL;
		parts = NULL;
	}

	/* We used to remount the entire tmpfs readonly if any :ro or
	 * :mixed mode was specified. However, Ubuntu's mountall has the
	 * unfortunate behavior to block bootup if /sys/fs/cgroup is
	 * mounted read-only and cannot be remounted read-write.
	 * (mountall reads /lib/init/fstab and tries to (re-)mount all of
	 * these if they are not already mounted with the right options;
	 * it contains an entry for /sys/fs/cgroup. In case it can't do
	 * that, it prompts for the user to either manually fix it or
	 * boot anyway. But without user input, booting of the container
	 * hangs.)
	 *
	 * Instead of remounting the entire tmpfs readonly, we only
	 * remount the paths readonly that are part of the cgroup
	 * hierarchy.
	 */

	free(path);

	return true;

out_error:
	saved_errno = errno;
	free(path);
	free(dirname);
	free(parts);
	free(abs_path);
	free(abs_path2);
	errno = saved_errno;
	return false;
}

static int cgfs_nrtasks(void *hdata)
{
	struct cgfs_data *d = hdata;
	struct cgroup_process_info *info;
	struct cgroup_mount_point *mp = NULL;
	char *abs_path = NULL;
	int ret;

	if (!d) {
		errno = ENOENT;
		return -1;
	}

	info = d->info;
	if (!info) {
		errno = ENOENT;
		return -1;
	}

	if (info->designated_mount_point) {
		mp = info->designated_mount_point;
	} else {
		mp = lxc_cgroup_find_mount_point(info->hierarchy, info->cgroup_path, false);
		if (!mp)
			return -1;
	}

	abs_path = cgroup_to_absolute_path(mp, info->cgroup_path, NULL);
	if (!abs_path)
		return -1;

	ret = cgroup_recursive_task_count(abs_path);
	free(abs_path);
	return ret;
}

static struct cgroup_process_info *
lxc_cgroup_process_info_getx(const char *proc_pid_cgroup_str,
			     struct cgroup_meta_data *meta)
{
	struct cgroup_process_info *result = NULL;
	FILE *proc_pid_cgroup = NULL;
	char *line = NULL;
	size_t sz = 0;
	int saved_errno = 0;
	struct cgroup_process_info **cptr = &result;
	struct cgroup_process_info *entry = NULL;

	proc_pid_cgroup = fopen_cloexec(proc_pid_cgroup_str, "r");
	if (!proc_pid_cgroup)
		return NULL;

	while (getline(&line, &sz, proc_pid_cgroup) != -1) {
		/* file format: hierarchy:subsystems:group */
		char *colon1;
		char *colon2;
		char *endptr;
		int hierarchy_number;
		struct cgroup_hierarchy *h = NULL;

		if (!line[0])
			continue;

		if (line[strlen(line) - 1] == '\n')
			line[strlen(line) - 1] = '\0';

		colon1 = strchr(line, ':');
		if (!colon1)
			continue;
		*colon1++ = '\0';
		colon2 = strchr(colon1, ':');
		if (!colon2)
			continue;
		*colon2++ = '\0';

		endptr = NULL;

		/* With cgroupv2 /proc/self/cgroup can contain entries of the
		 * form: 0::/
		 * These entries need to be skipped.
		 */
		if (!strcmp(colon1, ""))
			continue;

		hierarchy_number = strtoul(line, &endptr, 10);
		if (!endptr || *endptr)
			continue;

		if (hierarchy_number > meta->maximum_hierarchy) {
			/* we encountered a hierarchy we didn't have before,
			 * so probably somebody remounted some stuff in the
			 * mean time...
			 */
			errno = EAGAIN;
			goto out_error;
		}

		h = meta->hierarchies[hierarchy_number];
		if (!h) {
			/* we encountered a hierarchy that was thought to be
			 * dead before, so probably somebody remounted some
			 * stuff in the mean time...
			 */
			errno = EAGAIN;
			goto out_error;
		}

		/* we are told that we should ignore this hierarchy */
		if (!h->used)
			continue;

		entry = calloc(1, sizeof(struct cgroup_process_info));
		if (!entry)
			goto out_error;

		entry->meta_ref = lxc_cgroup_get_meta(meta);
		entry->hierarchy = h;
		entry->cgroup_path = strdup(colon2);
		if (!entry->cgroup_path)
			goto out_error;
		prune_init_scope(entry->cgroup_path);

		*cptr = entry;
		cptr = &entry->next;
		entry = NULL;
	}

	fclose(proc_pid_cgroup);
	free(line);
	return result;

out_error:
	saved_errno = errno;
	if (proc_pid_cgroup)
		fclose(proc_pid_cgroup);
	lxc_cgroup_process_info_free(result);
	lxc_cgroup_process_info_free(entry);
	free(line);
	errno = saved_errno;
	return NULL;
}

static char **subsystems_from_mount_options(const char *mount_options,
					    char **kernel_list)
{
	char *token, *str, *saveptr = NULL;
	char **result = NULL;
	size_t result_capacity = 0;
	size_t result_count = 0;
	int saved_errno;
	int r;

	str = alloca(strlen(mount_options)+1);
	strcpy(str, mount_options);
	for (; (token = strtok_r(str, ",", &saveptr)); str = NULL) {
		/* we have a subsystem if it's either in the list of
		 * subsystems provided by the kernel OR if it starts
		 * with name= for named hierarchies
		 */
		r = lxc_grow_array((void ***)&result, &result_capacity, result_count + 1, 12);
		if (r < 0)
			goto out_free;
		result[result_count + 1] = NULL;
		if (strncmp(token, "name=", 5) && !lxc_string_in_array(token, (const char **)kernel_list)) {
			/* this is eg 'systemd' but the mount will be
			 * 'name=systemd'
			 */
			result[result_count] = malloc(strlen(token) + 6);
			if (result[result_count])
				sprintf(result[result_count], "name=%s", token);
		} else
			result[result_count] = strdup(token);
		if (!result[result_count])
			goto out_free;
		result_count++;
	}

	return result;

out_free:
	saved_errno = errno;
	lxc_free_array((void**)result, free);
	errno = saved_errno;
	return NULL;
}

static void lxc_cgroup_mount_point_free(struct cgroup_mount_point *mp)
{
	if (!mp)
		return;
	free(mp->mount_point);
	free(mp->mount_prefix);
	free(mp);
}

static void lxc_cgroup_hierarchy_free(struct cgroup_hierarchy *h)
{
	if (!h)
		return;
	if (h->subsystems) {
		lxc_free_array((void **)h->subsystems, free);
		h->subsystems = NULL;
	}
	if (h->all_mount_points) {
		free(h->all_mount_points);
		h->all_mount_points = NULL;
	}
	free(h);
	h = NULL;
}

static bool is_valid_cgroup(const char *name)
{
	const char *p;
	for (p = name; *p; p++) {
		/* Use the ASCII printable characters range(32 - 127)
		 * is reasonable, we kick out 32(SPACE) because it'll
		 * break legacy lxc-ls
		 */
		if (*p <= 32 || *p >= 127 || *p == '/')
			return false;
	}
	return strcmp(name, ".") != 0 && strcmp(name, "..") != 0;
}

static int create_or_remove_cgroup(bool do_remove,
		struct cgroup_mount_point *mp, const char *path, int recurse,
		struct lxc_conf *conf)
{
	int r, saved_errno = 0;
	char *buf = cgroup_to_absolute_path(mp, path, NULL);
	if (!buf)
		return -1;

	/* create or remove directory */
	if (do_remove) {
		if (!dir_exists(buf))
			return 0;
		if (recurse) {
			if (conf && !lxc_list_empty(&conf->id_map))
				r = userns_exec_1(conf, rmdir_wrapper, buf,
						  "rmdir_wrapper");
			else
				r = cgroup_rmdir(buf);
		} else
			r = rmdir(buf);
	} else
		r = mkdir_p(buf, 0777);
	saved_errno = errno;
	free(buf);
	errno = saved_errno;
	return r;
}

static int create_cgroup(struct cgroup_mount_point *mp, const char *path)
{
	return create_or_remove_cgroup(false, mp, path, false, NULL);
}

static int remove_cgroup(struct cgroup_mount_point *mp,
			 const char *path, bool recurse, struct lxc_conf *conf)
{
	return create_or_remove_cgroup(true, mp, path, recurse, conf);
}

static char *cgroup_to_absolute_path(struct cgroup_mount_point *mp,
				     const char *path, const char *suffix)
{
	/* first we have to make sure we subtract the mount point's prefix */
	char *prefix = mp->mount_prefix;
	char *buf;
	ssize_t len, rv;

	/* we want to make sure only absolute paths to cgroups are passed to us */
	if (path[0] != '/') {
		errno = EINVAL;
		return NULL;
	}

	if (prefix && !strcmp(prefix, "/"))
		prefix = NULL;

	/* prefix doesn't match */
	if (prefix && strncmp(prefix, path, strlen(prefix)) != 0) {
		errno = EINVAL;
		return NULL;
	}
	/* if prefix is /foo and path is /foobar */
	if (prefix && path[strlen(prefix)] != '/' && path[strlen(prefix)] != '\0') {
		errno = EINVAL;
		return NULL;
	}

	/* remove prefix from path */
	path += prefix ? strlen(prefix) : 0;

	len = strlen(mp->mount_point) + strlen(path) + (suffix ? strlen(suffix) : 0);
	buf = calloc(len + 1, 1);
	if (!buf)
		return NULL;
	rv = snprintf(buf, len + 1, "%s%s%s", mp->mount_point, path, suffix ? suffix : "");
	if (rv > len) {
		free(buf);
		errno = ENOMEM;
		return NULL;
	}

	return buf;
}

static struct cgroup_process_info *
find_info_for_subsystem(struct cgroup_process_info *info, const char *subsystem)
{
	struct cgroup_process_info *info_ptr;
	for (info_ptr = info; info_ptr; info_ptr = info_ptr->next) {
		struct cgroup_hierarchy *h = info_ptr->hierarchy;
		if (!h)
			continue;
		if (lxc_string_in_array(subsystem, (const char **)h->subsystems))
			return info_ptr;
	}
	errno = ENOENT;
	return NULL;
}

static int do_cgroup_get(const char *cgroup_path, const char *sub_filename,
			 char *value, size_t len)
{
	const char *parts[3] = {
		cgroup_path,
		sub_filename,
		NULL
	};
	char *filename;
	int ret, saved_errno;

	filename = lxc_string_join("/", parts, false);
	if (!filename)
		return -1;

	ret = lxc_read_from_file(filename, value, len);
	saved_errno = errno;
	free(filename);
	errno = saved_errno;
	return ret;
}

static int do_cgroup_set(const char *cgroup_path, const char *sub_filename,
			 const char *value)
{
	const char *parts[3] = {
		cgroup_path,
		sub_filename,
		NULL
	};
	char *filename;
	int ret, saved_errno;

	filename = lxc_string_join("/", parts, false);
	if (!filename)
		return -1;

	ret = lxc_write_to_file(filename, value, strlen(value), false);
	saved_errno = errno;
	free(filename);
	errno = saved_errno;
	return ret;
}

static int do_setup_cgroup_limits(struct cgfs_data *d,
			   struct lxc_list *cgroup_settings, bool do_devices)
{
	struct lxc_list *iterator, *sorted_cgroup_settings, *next;
	struct lxc_cgroup *cg;
	int ret = -1;

	if (lxc_list_empty(cgroup_settings))
		return 0;

	sorted_cgroup_settings = sort_cgroup_settings(cgroup_settings);
	if (!sorted_cgroup_settings) {
		return -1;
	}

	lxc_list_for_each(iterator, sorted_cgroup_settings) {
		cg = iterator->elem;

		if (do_devices == !strncmp("devices", cg->subsystem, 7)) {
			if (strcmp(cg->subsystem, "devices.deny") == 0 &&
					cgroup_devices_has_allow_or_deny(d, cg->value, false))
				continue;
			if (strcmp(cg->subsystem, "devices.allow") == 0 &&
					cgroup_devices_has_allow_or_deny(d, cg->value, true))
				continue;
			if (lxc_cgroup_set_data(cg->subsystem, cg->value, d)) {
				if (do_devices && (errno == EACCES || errno == EPERM)) {
					WARN("Error setting %s to %s for %s",
					      cg->subsystem, cg->value, d->name);
					continue;
				}
				SYSERROR("Error setting %s to %s for %s",
				      cg->subsystem, cg->value, d->name);
				goto out;
			}
		}

		DEBUG("cgroup '%s' set to '%s'", cg->subsystem, cg->value);
	}

	ret = 0;
	INFO("cgroup has been setup");
out:
	lxc_list_for_each_safe(iterator, sorted_cgroup_settings, next) {
		lxc_list_del(iterator);
		free(iterator);
	}
	free(sorted_cgroup_settings);
	return ret;
}

static bool cgroup_devices_has_allow_or_deny(struct cgfs_data *d,
					     char *v, bool for_allow)
{
	char *path;
	FILE *devices_list;
	char *line = NULL;
	size_t sz = 0;
	bool ret = !for_allow;
	const char *parts[3] = {
		NULL,
		"devices.list",
		NULL
	};

	/* XXX FIXME if users could use something other than 'lxc.devices.deny =
	 * a'.  not sure they ever do, but they *could* right now, I'm assuming
	 * they do NOT
	 */
	if (!for_allow && strcmp(v, "a") != 0 && strcmp(v, "a *:* rwm") != 0)
		return false;

	parts[0] = (const char *)lxc_cgroup_get_hierarchy_abs_path_data("devices", d);
	if (!parts[0])
		return false;
	path = lxc_string_join("/", parts, false);
	if (!path) {
		free((void *)parts[0]);
		return false;
	}

	devices_list = fopen_cloexec(path, "r");
	if (!devices_list) {
		free(path);
		return false;
	}

	while (getline(&line, &sz, devices_list) != -1) {
		size_t len = strlen(line);
		if (len > 0 && line[len-1] == '\n')
			line[len-1] = '\0';
		if (strcmp(line, "a *:* rwm") == 0) {
			ret = for_allow;
			goto out;
		} else if (for_allow && strcmp(line, v) == 0) {
			ret = true;
			goto out;
		}
	}

out:
	fclose(devices_list);
	free(line);
	free(path);
	return ret;
}

static int cgroup_recursive_task_count(const char *cgroup_path)
{
	DIR *d;
	struct dirent *dent;
	int n = 0, r;

	d = opendir(cgroup_path);
	if (!d)
		return 0;

	while ((dent = readdir(d))) {
		const char *parts[3] = {
			cgroup_path,
			dent->d_name,
			NULL
		};
		char *sub_path;
		struct stat st;

		if (!strcmp(dent->d_name, ".") || !strcmp(dent->d_name, ".."))
			continue;
		sub_path = lxc_string_join("/", parts, false);
		if (!sub_path) {
			closedir(d);
			return -1;
		}
		r = stat(sub_path, &st);
		if (r < 0) {
			closedir(d);
			free(sub_path);
			return -1;
		}
		if (S_ISDIR(st.st_mode)) {
			r = cgroup_recursive_task_count(sub_path);
			if (r >= 0)
				n += r;
		} else if (!strcmp(dent->d_name, "tasks")) {
			r = lxc_count_file_lines(sub_path);
			if (r >= 0)
				n += r;
		}
		free(sub_path);
	}
	closedir(d);

	return n;
}

static int handle_cgroup_settings(struct cgroup_mount_point *mp,
				  char *cgroup_path)
{
	int r, saved_errno = 0;
	char buf[2];

	mp->need_cpuset_init = false;

	/* If this is the memory cgroup, we want to enforce hierarchy.
	 * But don't fail if for some reason we can't.
	 */
	if (lxc_string_in_array("memory", (const char **)mp->hierarchy->subsystems)) {
		char *cc_path = cgroup_to_absolute_path(mp, cgroup_path, "/memory.use_hierarchy");
		if (cc_path) {
			r = lxc_read_from_file(cc_path, buf, 1);
			if (r < 1 || buf[0] != '1') {
				r = lxc_write_to_file(cc_path, "1", 1, false);
				if (r < 0)
					SYSERROR("failed to set memory.use_hierarchy to 1; continuing");
			}
			free(cc_path);
		}
	}

	/* if this is a cpuset hierarchy, we have to set cgroup.clone_children in
	 * the base cgroup, otherwise containers will start with an empty cpuset.mems
	 * and cpuset.cpus and then
	 */
	if (lxc_string_in_array("cpuset", (const char **)mp->hierarchy->subsystems)) {
		char *cc_path = cgroup_to_absolute_path(mp, cgroup_path, "/cgroup.clone_children");
		struct stat sb;

		if (!cc_path)
			return -1;
		/* cgroup.clone_children is not available when running under
		 * older kernel versions; in this case, we'll initialize
		 * cpuset.cpus and cpuset.mems later, after the new cgroup
		 * was created
		 */
		if (stat(cc_path, &sb) != 0 && errno == ENOENT) {
			mp->need_cpuset_init = true;
			free(cc_path);
			return 0;
		}
		r = lxc_read_from_file(cc_path, buf, 1);
		if (r == 1 && buf[0] == '1') {
			free(cc_path);
			return 0;
		}
		r = lxc_write_to_file(cc_path, "1", 1, false);
		saved_errno = errno;
		free(cc_path);
		errno = saved_errno;
		return r < 0 ? -1 : 0;
	}
	return 0;
}

static int cgroup_read_from_file(const char *fn, char buf[], size_t bufsize)
{
	int ret = lxc_read_from_file(fn, buf, bufsize);
	if (ret < 0) {
		SYSERROR("failed to read %s", fn);
		return ret;
	}
	if (ret == bufsize) {
		if (bufsize > 0) {
			/* obviously this wasn't empty */
			buf[bufsize-1] = '\0';
			return ret;
		}
		/* Callers don't do this, but regression/sanity check */
		ERROR("was not expecting 0 bufsize");
		return -1;
	}
	buf[ret] = '\0';
	return ret;
}

static bool do_init_cpuset_file(struct cgroup_mount_point *mp,
				const char *path, const char *name)
{
	char value[1024];
	char *childfile, *parentfile = NULL, *tmp;
	int ret;
	bool ok = false;

	childfile = cgroup_to_absolute_path(mp, path, name);
	if (!childfile)
		return false;

	/* don't overwrite a non-empty value in the file */
	ret = cgroup_read_from_file(childfile, value, sizeof(value));
	if (ret < 0)
		goto out;
	if (value[0] != '\0' && value[0] != '\n') {
		ok = true;
		goto out;
	}

	/* path to the same name in the parent cgroup */
	parentfile = strdup(path);
	if (!parentfile)
		goto out;

	tmp = strrchr(parentfile, '/');
	if (!tmp)
		goto out;
	if (tmp == parentfile)
		tmp++; /* keep the '/' at the start */
	*tmp = '\0';
	tmp = parentfile;
	parentfile = cgroup_to_absolute_path(mp, tmp, name);
	free(tmp);
	if (!parentfile)
		goto out;

	/* copy from parent to child cgroup */
	ret = cgroup_read_from_file(parentfile, value, sizeof(value));
	if (ret < 0)
		goto out;
	if (ret == sizeof(value)) {
		/* If anyone actually sees this error, we can address it */
		ERROR("parent cpuset value too long");
		goto out;
	}
	ok = (lxc_write_to_file(childfile, value, strlen(value), false) >= 0);
	if (!ok)
		SYSERROR("failed writing %s", childfile);

out:
	free(parentfile);
	free(childfile);
	return ok;
}

static bool init_cpuset_if_needed(struct cgroup_mount_point *mp,
				  const char *path)
{
	/* the files we have to handle here are only in cpuset hierarchies */
	if (!lxc_string_in_array("cpuset",
				 (const char **)mp->hierarchy->subsystems))
		return true;

	if (!mp->need_cpuset_init)
		return true;

	return (do_init_cpuset_file(mp, path, "/cpuset.cpus") &&
		do_init_cpuset_file(mp, path, "/cpuset.mems") );
}

static void print_cgfs_init_debuginfo(struct cgfs_data *d)
{
	int i;

	if (!getenv("LXC_DEBUG_CGFS"))
		return;

	DEBUG("Cgroup information:");
	DEBUG("  container name: %s", d->name);
	if (!d->meta || !d->meta->hierarchies) {
		DEBUG("  No hierarchies found.");
		return;
	}
	DEBUG("  Controllers:");
	for (i = 0; i <= d->meta->maximum_hierarchy; i++) {
		char **p;
		struct cgroup_hierarchy *h = d->meta->hierarchies[i];
		if (!h) {
			DEBUG("     Empty hierarchy number %d.", i);
			continue;
		}
		for (p = h->subsystems; p && *p; p++) {
			DEBUG("     %2d: %s", i, *p);
		}
	}
}

struct cgroup_ops *cgfs_ops_init(void)
{
	return &cgfs_ops;
}

static void *cgfs_init(struct lxc_handler *handler)
{
	struct cgfs_data *d;

	d = malloc(sizeof(*d));
	if (!d)
		return NULL;

	memset(d, 0, sizeof(*d));
	d->name = strdup(handler->name);
	if (!d->name)
		goto err1;

	d->cgroup_pattern = lxc_global_config_value("lxc.cgroup.pattern");

	d->meta = lxc_cgroup_load_meta();
	if (!d->meta) {
		ERROR("cgroupfs failed to detect cgroup metadata");
		goto err2;
	}

	print_cgfs_init_debuginfo(d);

	return d;

err2:
	free(d->name);
err1:
	free(d);
	return NULL;
}

static void cgfs_destroy(void *hdata, struct lxc_conf *conf)
{
	struct cgfs_data *d = hdata;

	if (!d)
		return;
	free(d->name);
	lxc_cgroup_process_info_free_and_remove(d->info, conf);
	lxc_cgroup_put_meta(d->meta);
	free(d);
}

static inline bool cgfs_create(void *hdata)
{
	struct cgfs_data *d = hdata;
	struct cgroup_process_info *i;
	struct cgroup_meta_data *md;

	if (!d)
		return false;
	md = d->meta;
	i = lxc_cgroupfs_create(d->name, d->cgroup_pattern, md, NULL);
	if (!i)
		return false;
	d->info = i;
	return true;
}

static inline bool cgfs_enter(void *hdata, pid_t pid)
{
	struct cgfs_data *d = hdata;
	struct cgroup_process_info *i;
	int ret;

	if (!d)
		return false;
	i = d->info;
	ret = lxc_cgroupfs_enter(i, pid, false);

	return ret == 0;
}

static inline bool cgfs_create_legacy(void *hdata, pid_t pid)
{
	struct cgfs_data *d = hdata;
	struct cgroup_process_info *i;

	if (!d)
		return false;
	i = d->info;
	if (lxc_cgroup_create_legacy(i, d->name, pid) < 0) {
		ERROR("failed to create legacy ns cgroups for '%s'", d->name);
		return false;
	}
	return true;
}

static const char *cgfs_get_cgroup(void *hdata, const char *subsystem)
{
	struct cgfs_data *d = hdata;

	if (!d)
		return NULL;
	return lxc_cgroup_get_hierarchy_path_data(subsystem, d);
}

static bool cgfs_escape(void *hdata)
{
	struct cgroup_meta_data *md;
	int i;
	bool ret = false;

	md = lxc_cgroup_load_meta();
	if (!md)
		return false;

	for (i = 0; i <= md->maximum_hierarchy; i++) {
		struct cgroup_hierarchy *h = md->hierarchies[i];
		struct cgroup_mount_point *mp;
		char *tasks;
		FILE *f;
		int written;

		if (!h) {
			WARN("not escaping hierarchy %d", i);
			continue;
		}

		mp = lxc_cgroup_find_mount_point(h, "/", true);
		if (!mp)
			goto out;

		tasks = cgroup_to_absolute_path(mp, "/", "tasks");
		if (!tasks)
			goto out;

		f = fopen(tasks, "a");
		free(tasks);
		if (!f)
			goto out;

		written = fprintf(f, "%d\n", lxc_raw_getpid());
		fclose(f);
		if (written < 0) {
			SYSERROR("writing tasks failed\n");
			goto out;
		}
	}

	ret = true;
out:
	lxc_cgroup_put_meta(md);
	return ret;
}

static int cgfs_num_hierarchies(void)
{
	/* not implemented */
	return -1;
}

static bool cgfs_get_hierarchies(int i, char ***out)
{
	/* not implemented */
	return false;
}

static bool cgfs_unfreeze(void *hdata)
{
	struct cgfs_data *d = hdata;
	char *cgabspath, *cgrelpath;
	int ret;

	if (!d)
		return false;

	cgrelpath = lxc_cgroup_get_hierarchy_path_data("freezer", d);
	cgabspath = lxc_cgroup_find_abs_path("freezer", cgrelpath, true, NULL);
	if (!cgabspath)
		return false;

	ret = do_cgroup_set(cgabspath, "freezer.state", "THAWED");
	free(cgabspath);
	return ret == 0;
}

static bool cgroupfs_setup_limits(void *hdata, struct lxc_list *cgroup_conf,
				  bool with_devices)
{
	struct cgfs_data *d = hdata;

	if (!d)
		return false;
	return do_setup_cgroup_limits(d, cgroup_conf, with_devices) == 0;
}

static bool lxc_cgroupfs_attach(const char *name, const char *lxcpath, pid_t pid)
{
	struct cgroup_meta_data *meta_data;
	struct cgroup_process_info *container_info;
	int ret;

	meta_data = lxc_cgroup_load_meta();
	if (!meta_data) {
		ERROR("could not move attached process %d to cgroup of container", pid);
		return false;
	}

	container_info = lxc_cgroup_get_container_info(name, lxcpath, meta_data);
	lxc_cgroup_put_meta(meta_data);
	if (!container_info) {
		ERROR("could not move attached process %d to cgroup of container", pid);
		return false;
	}

	ret = lxc_cgroupfs_enter(container_info, pid, false);
	lxc_cgroup_process_info_free(container_info);
	if (ret < 0) {
		ERROR("could not move attached process %d to cgroup of container", pid);
		return false;
	}
	return true;
}

struct chown_data {
	const char *cgroup_path;
	uid_t origuid;
};

/*
 * TODO - someone should refactor this to unshare once passing all the paths
 * to be chowned in one go
 */
static int chown_cgroup_wrapper(void *data)
{
	struct chown_data *arg = data;
	uid_t destuid;
	char *fpath;

	if (setresgid(0,0,0) < 0)
		SYSERROR("Failed to setgid to 0");
	if (setresuid(0,0,0) < 0)
		SYSERROR("Failed to setuid to 0");
	if (setgroups(0, NULL) < 0)
		SYSERROR("Failed to clear groups");
	destuid = get_ns_uid(arg->origuid);

	if (chown(arg->cgroup_path, destuid, 0) < 0)
		SYSERROR("Failed chowning %s to %d", arg->cgroup_path, (int)destuid);

	fpath = lxc_append_paths(arg->cgroup_path, "tasks");
	if (!fpath)
		return -1;
	if (chown(fpath, destuid, 0) < 0)
		SYSERROR("Error chowning %s\n", fpath);
	free(fpath);

	fpath = lxc_append_paths(arg->cgroup_path, "cgroup.procs");
	if (!fpath)
		return -1;
	if (chown(fpath, destuid, 0) < 0)
		SYSERROR("Error chowning %s", fpath);
	free(fpath);

	return 0;
}

static bool do_cgfs_chown(char *cgroup_path, struct lxc_conf *conf)
{
	struct chown_data data;
	char *fpath;

	if (!dir_exists(cgroup_path))
		return true;

	if (lxc_list_empty(&conf->id_map))
		/* If there's no mapping then we don't need to chown */
		return true;

	data.cgroup_path = cgroup_path;
	data.origuid = geteuid();

	/* Unpriv users can't chown it themselves, so chown from
	 * a child namespace mapping both our own and the target uid
	 */
	if (userns_exec_1(conf, chown_cgroup_wrapper, &data,
			  "chown_cgroup_wrapper") < 0) {
		ERROR("Error requesting cgroup chown in new namespace");
		return false;
	}

	/*
	 * Now chmod 775 the directory else the container cannot create cgroups.
	 * This can't be done in the child namespace because it only group-owns
	 * the cgroup
	 */
	if (chmod(cgroup_path, 0775) < 0) {
		SYSERROR("Error chmoding %s\n", cgroup_path);
		return false;
	}
	fpath = lxc_append_paths(cgroup_path, "tasks");
	if (!fpath)
		return false;
	if (chmod(fpath, 0664) < 0)
		SYSERROR("Error chmoding %s\n", fpath);
	free(fpath);
	fpath = lxc_append_paths(cgroup_path, "cgroup.procs");
	if (!fpath)
		return false;
	if (chmod(fpath, 0664) < 0)
		SYSERROR("Error chmoding %s\n", fpath);
	free(fpath);

	return true;
}

static bool cgfs_chown(void *hdata, struct lxc_conf *conf)
{
	struct cgfs_data *d = hdata;
	struct cgroup_process_info *info_ptr;
	char *cgpath;
	bool r = true;

	if (!d)
		return false;

	for (info_ptr = d->info; info_ptr; info_ptr = info_ptr->next) {
		if (!info_ptr->hierarchy)
			continue;

		if (!info_ptr->designated_mount_point) {
			info_ptr->designated_mount_point = lxc_cgroup_find_mount_point(info_ptr->hierarchy, info_ptr->cgroup_path, true);
			if (!info_ptr->designated_mount_point) {
				SYSERROR("Could not chown cgroup %s: internal error (couldn't find any writable mountpoint to cgroup filesystem)", info_ptr->cgroup_path);
				return false;
			}
		}

		cgpath = cgroup_to_absolute_path(info_ptr->designated_mount_point, info_ptr->cgroup_path, NULL);
		if (!cgpath) {
			SYSERROR("Could not chown cgroup %s: internal error", info_ptr->cgroup_path);
			continue;
		}
		r = do_cgfs_chown(cgpath, conf);
		if (!r && is_crucial_hierarchy(info_ptr->hierarchy)) {
			ERROR("Failed chowning %s\n", cgpath);
			free(cgpath);
			return false;
		}
		free(cgpath);
	}

	return true;
}

static struct cgroup_ops cgfs_ops = {
	.init = cgfs_init,
	.destroy = cgfs_destroy,
	.create = cgfs_create,
	.enter = cgfs_enter,
	.create_legacy = cgfs_create_legacy,
	.get_cgroup = cgfs_get_cgroup,
	.escape = cgfs_escape,
	.num_hierarchies = cgfs_num_hierarchies,
	.get_hierarchies = cgfs_get_hierarchies,
	.get = lxc_cgroupfs_get,
	.set = lxc_cgroupfs_set,
	.unfreeze = cgfs_unfreeze,
	.setup_limits = cgroupfs_setup_limits,
	.name = "cgroupfs",
	.attach = lxc_cgroupfs_attach,
	.chown = cgfs_chown,
	.mount_cgroup = cgroupfs_mount_cgroup,
	.nrtasks = cgfs_nrtasks,
	.driver = CGFS,
};